HE4 is a novel biomarker that complements CA125 measurement in patients with ovarian cancer by providing improved sensitivity at fixed levels of specificity. This combination has been used to develop the ROMA algorithm (Risk of Ovarian Malignancy Algorithm) that will improve triage of women with adnexal mass.
by Dr W. J. Allard and Dr R. G. Moore
Human epididymis protein 4 (HE4) is a low molecular weight (25 Kd) member of the Whey Acidic Protein family of protease inhibitors. Because it contains two of the 4-disulphide core domains characteristic of this family, it is sometimes referred to as whey acidic protein four disulphide core protein 2, WFDC2 [1]. No substrate or function has been ascribed to this protein although it is likely to be secreted and is expressed primarily in epithelia of normal female genital tissues such as fallopian tube, endometrium and endocervix, and in the respiratory epithelium, particularly in the trachea [2]. HE4 gene expression is up-regulated in invasive epithelial ovarian cancer as well as adenocarcinomas of the lung.
Ovarian cancer
Ovarian cancer is diagnosed annually in more than 200,000 women worldwide, with the greatest incidence in the US and Northern Europe, and lowest incidence in Africa and Asia [3]. Ovarian cancer is the fourth leading cause of cancer death worldwide and is responsible for 5% of all cancer deaths in women. Approximately 1 in every 57 women in the US will die of this disease. With the use of mammography in developed countries, over 90% of breast cancers are diagnosed at an early and potentially curable stage. However, fewer than 30% of all ovarian cancers are diagnosed in Stages I/II and worldwide mortality from ovarian cancer has decreased only 12% since 1973.
These grim statistics reflect the lack of effective tools for early detection of ovarian cancer. Mammography and cervical cancer screening with PAP smears have led to a shift in the diagnosis of breast and cervical cancer to earlier stages and to pre-invasive disease that are fundamentally curable, leading to a decrease in mortality for these cancers. However, no such tools are available for early detection of ovarian cancer. While the serum CA125 blood test is effective for monitoring women with ovarian cancer for progression or recurrence, the sensitivity and specificity are both approximately 80%, which is too low for effective screening [4].
There are three approaches that could lead to improved survival for women with ovarian cancer. These include improvements in therapy, better screening strategies to identify ovarian cancer at early and potentially treatable stages, and improved patient management. The implementation of platinum and taxol-based chemotherapies in the 1980s provided improved survival that has been enhanced by intraperitoneal delivery. Although early detection strategies have not proven successful so far, novel approaches to patient management using biomarkers and statistical algorithms have been suggested.
Ovarian cancer is often discovered on surgery for an adnexal mass. Many women with adnexal masses present with symptoms, and recent studies have shown that women with ovarian cancer have symptoms that may distinguish them from women with benign tumours. These symptoms include bloating, pelvic and abdominal pain, urinary urgency and difficulty eating, or feeling full. While these symptoms were also present in women with benign diseases, in women with ovarian cancer the symptoms were more persistent (> 12 times per month) and were of recent origin (< 1 year) [5]. The management of women with ovarian cancer can be improved through more effective triage of symptomatic women to appropriate specialists at experienced medical centres. A number of studies have demonstrated that women with ovarian cancer have decreased morbidity and improved survival if their surgery is performed by a specialist in gynaecological oncology and at institutions with multidisciplinary teams experienced in the management of ovarian cancer [6]. Unfortunately, not all women with ovarian cancer have their surgery performed by specialists in gynaecological oncology, who are the specialists trained in the management of ovarian cancer and the surgeons that perform adequate surgical staging and cytoreductive surgery most frequently. In the US, for example, fewer than half of women with ovarian cancer have their surgery performed by a specialist in gynaecological oncology. It is important therefore, to develop better diagnostic approaches so that women with ovarian cancer will be referred to oncology specialists in high volume institutions.
ROMA algorithm for triage of women presenting with adnexal mass
One approach to improving the triage of women with an adnexal mass is to combine multiple biomarkers to improve sensitivity. In a recent study, nine biomarkers were evaluated, in sera of women with an adnexal mass, for their ability to improve the sensitivity of CA125 for prediction of the presence of ovarian cancer. Logistic regression and cross validation were performed at predetermined specificities of 90%, 95% and 98%. Of the nine biomarkers studied, only HE4 significantly improved the sensitivity of CA125 [7]. For all patients, the combination of CA125 and HE4 provided 80.7% sensitivity at 90% specificity, compared with 61.2% and 77.6% for CA125 and HE4 alone, respectively. The improvement was particularly striking in women with Stage I disease, where HE4 and CA125 combined provided 46.1% sensitivity at 90% specificity, while the sensitivities of CA125 and HE4 alone were 23.1% and 46.2%. This finding suggests that HE4 has increased sensitivity in early stage disease over that of CA125.
CA125 and HE4 were combined in a logistic regression model to stratify women presenting with adnexal mass into groups with low and high risk, and this model was validated in a prospective, multicentre double blind trial of 503 evaluable patients scheduled for surgery for adnexal mass [8]. A total of 129 epithelial ovarian cancers and 22 tumours of low malignant potential were evaluated. The predictive algorithm, termed Risk of Ovarian Malignancy Algorithm (ROMA), was validated prospectively by setting specificity at 75% and determining sensitivity. On examining all women in the study with benign disease, epithelial ovarian cancer (EOC) and tumours of low malignant potential (LMP tumours), the authors reported a sensitivity of 92% in postmenopausal women, 76% in premenopausal women and 89% in pre and post-menopausal women combined. Further analysis of patients with invasive epithelial ovarian cancer showed that nearly 95% of postmenopausal and 89% of premenopausal women with EOC were correctly identified. Among the 17 patients with an EOC or LMP that were classified into the low risk subgroup, more than half had LMP tumours, and only one patient with a Stage III ovarian cancer was misclassified into the low risk subgroup.
There is no single current standard of care for the risk assessment and triage of women presenting with an adnexal mass or ovarian cyst. Various protocols have been suggested such as the recommendations from the American College of Obstetrics and Gynecology (ACOG) or the Society of Gynecologic Oncologists (SGO). Each of these clinical guidelines employs the use of history, physical exam, family history, imaging and often CA125. However, these protocols rely on subjective data that do not give a quantitative risk calculation. Jacobs et al. developed an algorithm, the Risk of Malignancy Index (RMI), that multiplies the concentration of CA125 (U/mL) X menopausal status (1 if premenopausal and 3 if postmenopausal) X ultrasound findings (scored as 0, 1 or 3) to produce a quantitative risk assessment [9]. The RMI has been validated in a number of separate trials and is currently being used in the clinical setting. Recently, at the Society of Gynecologic Oncologists annual meeting in 2009, Moore et al. presented their data on a direct comparison of the RMI with the ROMA at 75% specificity, and showed that in their study the ROMA significantly outperformed the RMI for all stages of disease [10].
If the ROMA algorithm were used to triage women with adnexal mass to the appropriate specialists, almost 90% of women with ovarian cancers and LMP tumours, and 95% of women with invasive epithelial ovarian cancers, would be referred to oncology specialists at high volume centres. Importantly, 75% of patients with non-malignant diseases would be spared referral to oncology specialists and could have their surgery performed safely within the community by a non-oncology specialist.
The use of HE4 for monitoring women with ovarian cancer
Early studies on CA 125 demonstrated that the sensitivity is approximately 80% at the Upper Limit of Normal of 35 U/mL [4]. The study by Moore et al. described above showed that CA125 + HE4 provided a sensitivity of 81%, compared to 61.2% for CA125 alone, at 90% specificity [7]. In addition, Brown et al. demonstrated that of the approximately 20% of ovarian cancer patients that had negative CA125 measurements, almost half had elevated concentrations of HE4 [11].
In a longitudinal study of 80 women under therapy for ovarian cancer or being monitored for recurrence, serial changes in CA125 and HE4 concentrations correlated with changes in clinical status in 79% of patients for CA125 compared with 76% of patients for HE4 [12]. Of the 17 patients with low CA125 levels, approximately 25% had interpretable changes in HE4 concentrations. As shown in Figure 1, the longitudinal changes in CA125 and HE4 are similar in the majority of patients and the results shown in Panel B are representative of these patients. However, for some patients, only HE4 will provide interpretable changes over time as shown in Panel A, and only CA 125 will provide useful longitudinal data in others. However, for all patients combined, CA 125 and HE4 provide clinically useful longitudinal changes in 84% of all patients.
Conclusions
HE4 represents the first novel biomarker that complements CA125 measurement in patients with ovarian cancer by providing improved sensitivity at fixed levels of specificity. This enhancement in sensitivity has been used to develop the ROMA algorithm that will improve triage of women with adnexal mass to the appropriate specialist and institution. In addition, the combination of CA125 and HE4 provides increased longitudinal sensitivity, leading to clinically interpretable biomarker data, for approximately half of the women being monitored for ovarian cancer with low and unchanging CA125 concentrations.
References
1. Bouchard D et al. Proteins with whey-acidic-protein motifs and cancer. Lancet Oncol 2006;7:167-74.
2. Drapkin R et al. Human epididymis protein 4 (HE4) is a secreted glycoprotein that is overexpressed by serous and endometrioid ovarian carcinomas. Cancer Res 2006;65:2162-69.
3. IARC. GLOBOCAN 2002. Cancer Incidence, Mortality and Prevalence Worldwide (2002 estimates). 2004.
4. Malkasian GD et al. Preoperative evaluation of premenopausal and postmenopausal patients with pelvic masses: Discrimination of benign from malignant disease. Am J Obstet Gynecol 1988;159:341-46.
5. Goff BA et al. Development of an ovarian cancer symptom index: possibilities for earlier detection. Cancer 2007;109:221-27.
6. Paulsen T et al. Improved short-term survival for advanced ovarian, tubal, and peritoneal cancer patients operated at teaching hospitals. Int J Gynecol Cancer 2006;16 Suppl 1:11-17.
7. Moore RG et al. The use of multiple novel tumor biomarkers for the detection of ovarian carcinoma in patients with a pelvic mass. Gynecologic Oncol 2008;108:402-408.
8. Moore RG et al. A novel multiple marker bioassay utilizing HE4 and CA 125 for the prediction of ovarian cancer in patients with a pelvic mass. Gynecologic Oncol 2009;112:40-46.
9. Jacobs I et al. A risk of malignancy index incorporating CA 125, ultrasound and menopausal status for the accurate preoperative diagnosis of ovarian cancer. Br J Obstet Gynecol 1990;97:922-29.
10. Moore RG et al. Comparison of a novel multiple marker assay versus the risk of malignancy index for the prediction of epithelial ovarian cancer in patients with a pelvic mass. Manuscript submitted.
11. Brown AK et al. Differential expression of CA 125 and a novel serum tumor marker HE4 in epithelial ovarian cancer. J Clin Oncol 2008;26 (Suppl):Abstract 5533.
12. Allard J et al. Use of a novel biomarker HE4 for monitoring patients with epithelial ovarian cancer. 2008;26 (Suppl):Abstract 5535.
The authors
W. Jeffrey Allard, Ph.D
Medivice Consulting, LLC
363 South Shore Road
New Durham, NH 03855, USA
jeff@mediviceconsulting.com
and Richard G. Moore, MD
Program in Women’s Oncology
Department of Obstetrics and Gynecology
Women and Infants’ Hospital
Alpert Medical School
Brown University
rmoore@wihri.org
